arXiv:2512.09871v1 Announce Type: new 
Abstract: Linear spectral mixture models (LMM) provide a concise form to disentangle the constituent materials (endmembers) and their corresponding proportions (abundance) in a single pixel. The critical challenges are how to model the spectral prior distribution and spectral variability. Prior knowledge and spectral variability can be rigorously modeled under the Bayesian framework, where posterior estimation of Abundance is derived by combining observed data with endmember prior distribution. Considering the key challenges and the advantages of the Bayesian framework, a novel method using a diffusion posterior sampler for semiblind unmixing, denoted as DPS4Un, is proposed to deal with these challenges with the following features: (1) we view the pretrained conditional spectrum diffusion model as a posterior sampler, which can combine the learned endmember prior with observation to get the refined abundance distribution. (2) Instead of using the existing spectral library as prior, which may raise bias, we establish the image-based endmember bundles within superpixels, which are used to train the endmember prior learner with diffusion model. Superpixels make sure the sub-scene is more homogeneous. (3) Instead of using the image-level data consistency constraint, the superpixel-based data fidelity term is proposed. (4) The endmember is initialized as Gaussian noise for each superpixel region, DPS4Un iteratively updates the abundance and endmember, contributing to spectral variability modeling. The experimental results on three real-world benchmark datasets demonstrate that DPS4Un outperforms the state-of-the-art hyperspectral unmixing methods.

تم اقتراح طريقة جديدة تُدعى DPS4Un لفك خلط البيانات الطيفية، باستخدام عينة خلفية للتشتت لنمذجة التباين الطيفي بشكل فعال وتحسين تقدير الوفرة. يدمج هذا النهج المعرفة السابقة ضمن إطار بايزي، مما يعالج التحديات الحرجة في نماذج الخلط الطيفي الخطي (LMM). تهدف الطريقة إلى تحسين توزيع الوفرة من خلال دمج المعلومات السابقة عن المواد مع البيانات الملاحظة.

Se ha propuesto un nuevo método llamado DPS4Un para el desmezclado hiperespectral, utilizando un muestreador posterior de difusión para modelar eficazmente la variabilidad espectral y mejorar la estimación de la abundancia. Este enfoque integra conocimientos previos dentro de un marco bayesiano, abordando desafíos críticos en los modelos de mezcla espectral lineales (LMM). El método busca refinar la distribución de abundancia combinando priors de endmembers aprendidos con datos observados.

Une nouvelle méthode appelée DPS4Un a été proposée pour le démasquage hyperspectral, utilisant un échantillonneur postérieur de diffusion pour modéliser efficacement la variabilité spectrale et améliorer l'estimation de l'abondance. Cette approche intègre des connaissances antérieures dans un cadre bayésien, abordant des défis critiques dans les modèles de mélange spectral linéaire (LMM). La méthode vise à affiner la distribution d'abondance en combinant des priors d'endmembers appris avec des données observées.

A novel method called DPS4Un has been proposed for hyperspectral unmixing, utilizing a diffusion posterior sampler to effectively model spectral variability and enhance abundance estimation. This approach integrates prior knowledge within a Bayesian framework, addressing critical challenges in linear spectral mixture models (LMM). The method aims to refine the abundance distribution by combining learned endmember priors with observed data.

Diffusion Posterior Sampler for Hyperspectral Unmixing with Spectral Variability Modeling

Was this article worth reading? Share it

LucidQuery AI

Supametas.AI

Attentive AI

LexiStock AI

SellerPic

Chroma

Ready to build your own newsroom?